问题分析

使用10046事件追踪方式，直接生成上传时的数据库事件日志进行分析，发现主要区别在于以下两条sql语句在每次长时间上传时都有出现，并且执行用户不是上传用户，而是数据库SYS用户。



********************************************************************************

SQL ID: frjd8zfy2jfdq Plan Hash: 510421217

SELECT executions, end_of_fetch_count,              elapsed_time/px_servers

  elapsed_time,        cpu_time/px_servers     cpu_time,

  buffer_gets/executions  buffer_gets

FROM

 (SELECT sum(executions)   as executions,                            sum(case

  when px_servers_executions > 0                              then

  px_servers_executions                                  else executions end)

  as px_servers,                sum(end_of_fetch_count) as end_of_fetch_count,

                sum(elapsed_time) as elapsed_time,

  sum(cpu_time)     as cpu_time,                     sum(buffer_gets)  as

  buffer_gets            FROM   gv$sql

  WHERE executions > 0                                 AND sql_id = :1

                              AND parsing_schema_name = :2)

call     count       cpu    elapsed       disk      query    current        rows

------- ------  -------- ---------- ---------- ---------- ----------  ----------

Parse       64      0.00       0.00          0          0          0           0

Execute     64      0.24       0.91          0          0          0           0

Fetch       64      0.14       0.42          0          0          0          64

call     count       cpu    elapsed       disk      query    current        rows

------- ------  -------- ---------- ---------- ---------- ----------  ----------

Parse       64      0.00       0.00          0          0          0           0

Execute     64      0.24       0.91          0          0          0           0

Fetch       64      0.14       0.42          0          0          0          64

------- ------  -------- ---------- ---------- ---------- ----------  ----------

total      192      0.39       1.34          0          0          0          64

Misses in library cache during parse: 1

Misses in library cache during execute: 1

Optimizer mode: CHOOSE

Parsing user id: SYS   (recursive depth: 1)

Number of plan statistics captured: 64

Rows (1st) Rows (avg) Rows (max)  Row Source Operation

---------- ---------- ----------  ---------------------------------------------------

         1          1          1  VIEW  (cr=0 pr=0 pw=0 time=20602 us)

         1          1          1   SORT AGGREGATE (cr=0 pr=0 pw=0 time=20596 us)

         0          0          0    PX COORDINATOR  (cr=0 pr=0 pw=0 time=20573 us)

         0          0          0     PX SEND QC (RANDOM) :TQ10000 (cr=0 pr=0 pw=0 time=0 us)

         0          0          0      VIEW  GV$SQL (cr=0 pr=0 pw=0 time=0 us)

         0          0          0       FIXED TABLE FIXED INDEX X$KGLCURSOR_CHILD (ind:2) (cr=0 pr=0 pw=0 time=0 us)

Elapsed times include waiting on following events: Event waited on                             Times   Max. Wait  Total Waited ----------------------------------------   Waited  ----------  ------------ PX Deq: reap credit                          1448        0.00          0.02

  PX Deq: Join ACK                              193        0.00          0.25

  IPC send completion sync                      128        0.00          0.08

  PX Deq: Parse Reply                           128        0.06          0.24

  PX Deq: Execute Reply                         128        0.00          0.06

  reliable message                               64        0.00          0.06

  PX Deq: Signal ACK EXT                        128        0.00          0.01

  PX Deq: Slave Session Stats                   128        0.00          0.01

  enq: PS - contention                           66        0.00          0.09

  KJC: Wait for msg sends to complete             5        0.00          0.00

  latch: shared pool                              1        0.00          0.00

********************************************************************************

SQL ID: 4b4wp0a8dvkf0 Plan Hash: 4033942373

SELECT executions, end_of_fetch_count,              elapsed_time/px_servers

  elapsed_time,        cpu_time/px_servers     cpu_time,

  buffer_gets/executions  buffer_gets

FROM

 (SELECT sum(executions_delta) as EXECUTIONS,

  sum(case when px_servers_execs_delta > 0

  then px_servers_execs_delta                                       else

  executions_delta end) as px_servers,

  sum(end_of_fetch_count_delta) as end_of_fetch_count,

  sum(elapsed_time_delta) as ELAPSED_TIME,              sum(cpu_time_delta)

  as CPU_TIME,                      sum(buffer_gets_delta) as BUFFER_GETS

       FROM   DBA_HIST_SQLSTAT s,

  V$DATABASE d,                                         DBA_HIST_SNAPSHOT sn

                           WHERE  s.dbid = d.dbid

       AND  bitand(nvl(s.flag, 0), 1) = 0                    AND

  sn.end_interval_time >                                   (select

  systimestamp at TIME ZONE dbtimezone                  from dual) - 7

                         AND  s.sql_id = :1

   AND  s.snap_id = sn.snap_id                           AND

  s.instance_number = sn.instance_number           AND  s.dbid = sn.dbid

                             AND  parsing_schema_name = :2)

call     count       cpu    elapsed       disk      query    current        rows

------- ------  -------- ---------- ---------- ---------- ----------  ----------

Parse       64      0.00       0.00          0          0          0           0

Execute     64      0.08       0.08          0          0          0           0

Fetch       64      0.13       0.27          0       1170          0          64

------- ------  -------- ---------- ---------- ---------- ----------  ----------

total      192      0.21       0.36          0       1170          0          64

Misses in library cache during parse: 1

Misses in library cache during execute: 1

Optimizer mode: CHOOSE

Parsing user id: SYS   (recursive depth: 1)

Number of plan statistics captured: 3

Rows (1st) Rows (avg) Rows (max)  Row Source Operation

---------- ---------- ----------  ---------------------------------------------------

         1          1          1  VIEW  (cr=18 pr=0 pw=0 time=6289 us cost=17 size=78 card=1)

         1          1          1   SORT AGGREGATE (cr=18 pr=0 pw=0 time=6282 us)

         0          0          0    NESTED LOOPS  (cr=18 pr=0 pw=0 time=6263 us cost=15 size=124 card=1)

         0          0          0     NESTED LOOPS  (cr=18 pr=0 pw=0 time=6258 us cost=15 size=124 card=1)

         0          0          0      HASH JOIN  (cr=18 pr=0 pw=0 time=6256 us cost=14 size=97 card=1)

         0          0          0       NESTED LOOPS  (cr=18 pr=0 pw=0 time=6244 us cost=14 size=97 card=1)

         0          0          0        STATISTICS COLLECTOR  (cr=18 pr=0 pw=0 time=6241 us)

         0          0          0         NESTED LOOPS  (cr=18 pr=0 pw=0 time=6223 us cost=13 size=81 card=1)

         1          1          1          MERGE JOIN CARTESIAN (cr=0 pr=0 pw=0 time=5648 us cost=0 size=10 card=1)

         1          1          1           FIXED TABLE FULL X$KCCDI (cr=0 pr=0 pw=0 time=3449 us cost=0 size=10 card=1)

         1          1          1           BUFFER SORT (cr=0 pr=0 pw=0 time=2187 us cost=0 size=0 card=1)

         1          1          1            FIXED TABLE FULL X$KCCDI2 (cr=0 pr=0 pw=0 time=2136 us cost=0 size=0 card=1)

         0          0          0          PARTITION RANGE ITERATOR PARTITION: KEY KEY (cr=18 pr=0 pw=0 time=568 us cost=13 size=71 card=1)

         0          0          0           TABLE ACCESS BY LOCAL INDEX ROWID BATCHED WRH$_SQLSTAT PARTITION: KEY KEY (cr=18 pr=0 pw=0 time=511 us cost=13 size=71 card=1)

         0          0          0            INDEX RANGE SCAN WRH$_SQLSTAT_INDEX PARTITION: KEY KEY (cr=18 pr=0 pw=0 time=477 us cost=11 size=0 card=2)(object id 8299)

         0          0          0        TABLE ACCESS BY INDEX ROWID WRM$_SNAPSHOT (cr=0 pr=0 pw=0 time=0 us cost=1 size=16 card=1)

         0          0          0         INDEX UNIQUE SCAN WRM$_SNAPSHOT_PK (cr=0 pr=0 pw=0 time=0 us cost=0 size=0 card=1)(object id 8695)

         0          0          0       TABLE ACCESS FULL WRM$_SNAPSHOT (cr=0 pr=0 pw=0 time=0 us cost=1 size=16 card=1)

         0          0          0      INDEX UNIQUE SCAN WRM$_SNAPSHOT_PK (cr=0 pr=0 pw=0 time=0 us cost=0 size=0 card=1)(object id 8695)

         0          0          0     TABLE ACCESS BY INDEX ROWID WRM$_SNAPSHOT (cr=0 pr=0 pw=0 time=0 us cost=1 size=27 card=1)

         0          0          0      FAST DUAL  (cr=0 pr=0 pw=0 time=0 us cost=2 size=0 card=1)

Elapsed times include waiting on following events:

  Event waited on                             Times   Max. Wait  Total Waited

  ----------------------------------------   Waited  ----------  ------------

  Disk file operations I/O                        3        0.00          0.00

  control file sequential read                  448        0.00          0.18

********************************************************************************

经过资料分析，这两条语句是由于开通了Oracle 12C的新功能而产生的，此功能为“optimizer_adaptive_features”。其主要功能是为了在语句执行过程中实时收集表的统计信息，方便Oracle选择更准确的执行计划。

但是，这个功能在RAC非常损耗性能，因为它需要查找全局视图gv$sql的数据，在多个实例的情况下会进行并发执行，参考这篇资料。

因为“optimizer_adaptive_features”参数宣称在OLAP数据仓库环境中可以获得较好的效果，实际在重上传轻查询的OLTP系统上，可以关闭这项新功能。

实验测试

因为该参数支持会话级别和系统级别调整，因为可以在单独一个会话内测试下插入时间和查询时间。

  • 插入测试

采用同时提交多条TARGETTABLE数据。

打开“optimizer_adaptive_features”时为2.2秒，关闭该功能时为0.375。

  • 查询测试

采用系统中执行最多的语句进行测试。

打开“optimizer_adaptive_features”时为0.05秒，关闭该功能时为0.02秒。

  • 关闭方式

因为“optimizer_adaptive_features”为动态参数，在系统级别修改时不需要重启。已在测试环境试验可用。

执行语句为

 alter system set optimizer_adaptive_features=false scope=both;

版权说明:camash原创，转载请注明出处 http://www.cnblogs.com/shenfeng/

--EOF--

Oracle 12C RAC的optimizer_adaptive_features造成数据插入超时的更多相关文章

  1. Oracle 12c RAC 静默安装文档

    参考文档: https://docs.oracle.com/en/database/oracle/oracle-database/12.2/cwlin/index.html https://docs. ...

  2. ORACLE 12c RAC的常用管理命令

    ORACLE 12c RAC的常用管理命令 一.查询 1.RAC查询 1.1.查询节点 [grid@swnode1 ~]$ su - grid [grid@swnode1 ~]$ /u01/app/1 ...

  3. Oracle 12c RAC 安装文档

    参考文档: https://docs.oracle.com/en/database/oracle/oracle-database/12.2/cwlin/index.html https://docs. ...

  4. 安装oracle 12c RAC遇到的一些问题

    (1) 安装grid软件,停止在38%很长时间不动,日志显示正常   解决方法: 由于是虚拟机安装,设置的内存为600M,关闭虚拟机,把内存调成1GB,问题解决~在38%Linking RMAN Ut ...

  5. oracle 12c RAC 重启

    转至:https://blog.csdn.net/weixin_40283570/article/details/81511072 关闭顺序 :关闭PDB----->关闭数据库------> ...

  6. Oracle 12c RAC 日志体系结构的变化

    1    说明 在11g中,查看GRID的日志,会进入$ORACLE_HOM/log. [grid@cndba.cn ~]$ cd $ORACLE_HOME/log/ [grid@cndba.cn l ...

  7. Oracle 12c RAC 搭建手册

    1  共享设备配置 1.1            设备划分说明 冗余策略 卷划分及大小说明 OCRVOTING Ocrvoting01 8G Ocrvoting02 8G Ocrvoting03 8G ...

  8. 问题：oracle 12c rac数据库服务器的home目录丢失问题解决2018-06-25 18:30

    问题原因:是由于运维粗心,在缩容/home(此目录下挂载了逻辑卷lv_home)时没有先缩小文件系统(resize2fs)也没有备份,导致home数据损坏,重启时系统无法正常启动 解决方案:跳过此ho ...

  9. Oracle 12c 新特性 --- 新增对数据泵操作的审计跟踪

    版权声明:本文为博主原创文章,未经博主允许不得转载. https://blog.csdn.net/leo__1990/article/details/90199263 概念 Oracle Data P ...

随机推荐

  1. SqlServer2008快照隔离模式的业务应用

    场景: 有200个检测点,每个检测点每天采集5个数据,对表的读写都是随机的(即有可能同时读写),总共有5年的数据. 存储方案A: 日期 点号 类型 值 20120101 001 A 1.0 20120 ...

  2. git 安装或者更新

    1. 安装编译git时需要的包 # yum install curl-devel expat-devel gettext-devel openssl-devel zlib-devel # yum in ...

  3. InLineHookSSDT

    //当Ring3调用OpenProcess //1从自己的模块(.exe)的导入表中取值 //2Ntdll.dll模块的导出表中执行ZwOpenProcess(取索引 进入Ring0层) //3进入R ...

  4. js正则表达式和replace

    javascript中replace与正则表达式 replace()最简单的算是能力就是简单的字符替换.示例代码如下: <script language="javascript&quo ...

  5. Redis系列-存储篇list主要操作函数小结

    在总结list之前,先要弄明白几个跟list相关的概念: 列表:一个从左到右的队列,个人理解更类似于一个栈,常规模式下,先进列表的元素,后出. 表头元素:列表最左端第一个元素. 表尾元素:列表最右端的 ...

  6. NOIP2013 提高组day2 2 花匠 动规 找拐点 树状数组

    花匠 描述 花匠栋栋种了一排花,每株花都有自己的高度.花儿越长越大,也越来越挤.栋栋决定把这排中的一部分花移走,将剩下的留在原地,使得剩下的花能有空间长大,同时,栋栋希望剩下的花排列得比较别致. 具体 ...

  7. wp8.1 C#技巧： Data和ViewModel类编写

    在Data.cs namespace PicApp { [DataContract] class DataItem : PropertyChangeNotification { public even ...

  8. Repeater控件的分页效果

    <webdiyer:AspNetPager ID="AspNetPager1" runat="server" HorizontalAlign=" ...

  9. SharePoint重置密码功能Demo

    博客地址 http://blog.csdn.net/foxdave 本文将说明一个简单的重置SharePoint用户密码(NTLM Windows认证)的功能如何实现 重置密码功能,实际上就是重置域用 ...

  10. java基础－008

     57.面向对象软件开发的优点 代码开发模块化,更易于维护 代码复用 增强代码的可靠性和灵活性 增强代码的可理解性 面向对象编程有很多重要的特性,比如:封装,继承,多态和抽象  58．封装 封装给对象 ...